Cationic Polymerization of Vinyl Monomers in Aqueous Media: From Monofunctional Oligomers to Long-Lived Polymer Chains SERGEI V. KOSTJUK†,§,* AND FRANCOIS GANACHAUD§ †
Research Institute for Physical Chemical Problems of the Belarusian State University, Leningradskaya St. 14, 220030 Minsk, Belarus, §Institut Charles Gerhardt, UMR5253 CNRS/UM2/ENSCM/UM1, Engineering of Macromolecular Architectures, ENSCM, 8 Rue de l’Ecole Normale, 34296 Montpellier Cedex, France RECEIVED ON JULY 10, 2009
CON SPECTUS
P
olymer latexes are easily prepared on a multimillion ton scale in industry using free radical initiated emulsion and suspension polymerizations in water, a cheap, nonviscous, heat-controlling, and environmentally benign solvent. Until recently, researchers had done little investigation into ionic polymerization because even a small amount of water would easily deactivate the conventional catalysts used in these processes. In the last decade, however, cationic polymerization in aqueous media has emerged as a new and attractive method for controlling the polymerization reactions using mild experimental conditions. This Account reviews the current science of and future outlook for cationic polymerization of vinyl monomers in aqueous media. We particularly emphasize the design and evolution of catalytic systems and the precision synthesis of functional polymers. Early work to tailor the suspension and emulsion cationic polymerizations of reactive monomers such as p-methoxystyrene and vinyl ethers used long-chain strong acids, called INISURF for their dual roles as initiators and surfactants, and lanthanide triflates. These polymerization processes shared two main features: (i) all reactions (initiation, propagation, and termination) occurred at the particle interface; (ii) synthesized polymers have limits on their molecular weight, attributed to the “critical DP” effect, related to the entry of oligomers inside the particles as they become increasingly hydrophobic. The next generation of catalysts, named “Lewis acid-surfactant combined catalysts” (LASC), shifted the polymerization locus from the interface to the inside of the monomer droplets, allowing for the production of long polymer chains. Recently, catalytic systems based on boranes, (BF3OEt2, B(C6F5)3, (C6F4-1,2-[B(C6F5)2]), and (C6F4-1,2-[B(C12F8)]2)), have shown great potential in controlling the cationic polymerization in “wet” solution, containing an excess of water relative to Lewis acid, or aqueous media of such industrially important monomers as styrene, cyclopentadiene, and even isobutylene.
Introduction
promising technique in the synthesis of well-de-
Since the discovery of the controlled/living cationic
fined homopolymers as well as different macro-
polymerization in the middle of the 1980s by the
molecular architectures using three main groups
1
2
Higashimura and Kennedy groups, considerable
of monomers: isobutylene, styrene and deriva-
attention has been paid to the utilization of this
tives, and vinyl ethers.3-7 A main drawback of
Published on the Web 12/03/2009 www.pubs.acs.org/acr 10.1021/ar900198q © 2010 American Chemical Society
Vol. 43, No. 3
March 2010
357-367
ACCOUNTS OF CHEMICAL RESEARCH
357
Cationic Polymerization of Vinyl Monomers in Aqueous Media Kostjuk and Ganachaud
the controlled cationic polymerization technique is the necessity to carry out the polymerization under strictly anhydrous conditions, since most conventional Lewis acids (LA) are decomposed in the presence of excess of water toward Lewis acid. Besides, the chlorinated solvents typically used in cationic polymerization are very toxic, and their utilization should be avoided. Ionic polymerization in aqueous media may seem like a contradiction at a first sight with respect to the high water sensitivity of commonly used polymerization catalysts, such as Lewis acids, alkyllithiums, and Grignard reagents. Such drawback may not appear as a deadlock, according to the variety of organometallic catalysts that have been developed during the last two decades for the (co)polymerization of olefins (ethylene, R-olefins, dienes, norbornenes) in aqueous media.8,9 Likewise, a number of catalytic systems were recently reported to promote the anionic polymerization of various cyclic and vinyl monomers in aqueous media (see refs 10-13 and references therein). Since the first report on the cationic polymerization of p-methoxystyrene (pMOS) in aqueous media using Yb(OTf)3 in 1999 by Sawamoto et al.,14 considerable advances in the field of cationic polymerization in aqueous media have emerged during the past ten years. This Account presents the most recent developments in this field according to the nature of catalysts used: (i) INISURFs (for initiator/surfactant); (ii) lanthanide triflates; (iii) Lewis acid-surfactant combined catalysts (LASC); (iv) BF3 · OEt2; (v) aromatic borane-based catalysts.
FIGURE 1. Simplified mechanism of the cationic polymerization in aqueous media [R ) H, 1-(4-methoxyphenyl)ethyl; A ) C12H25(C6H4)SO3-, LA · OH-; M ) monomer].
Another striking feature of ionic polymerization in aqueous media is the systematic limitation of the polymer molecular weights, in best cases set at about 3000 g/mol. This so-called “critical DP” effect is ascribed to the entry of oligomers inside the particles when they become too hydrophobic and lose their surface activity (Figure 1).10,11,15-17 In most cases, once the chains have penetrated inside the particles, they cannot undergo further propagation steps, since the catalyst is “locked” at the interface. We will show in this Account that recent works aimed (and sometimes succeeded) at driving the catalyst toward the core of the droplet/particle while keeping its Lewis acidity as high as possible to reactivate the oligomer chain-ends and produce living polymer chains.
General Mechanism of Ionic Polymerization in Aqueous Media
First Generation: Strong Brønsted Acid Catalysts
The main feature of almost all ionic polymerizations in aqueous media is the fact that all main reactions, that is, initiation, propagation, and termination, first (and often exclusively) proceed at the particle interface (Figure 1).10,11,15,16 Aqueous cationic polymerization is initiated by proton or carbocation derived from a strong/super acid or an initiator-Lewis acid pair, respectively. Propagation is very fast until water transfers to the growing carbocation to generate a nonpropagating chain Pn-OH (Figure 1). Small oligomers accumulate at the interface at the beginning of the reaction and decrease the water content at the interface. This effect is responsible for the observed increase in molecular weight with increasing monomer conversion10,15-17 and not a controlled polymerization process as first thought in early studies.14,18-20 In rare cases however, the termination reaction via water is reversible and the polymerization proceeds in a controlled fashion (vide infra).21
In a series of publications,16,22 dodecylbenzenesulfonic acid (DBSA) was used as an INISURF (initiator and surfactant combined) in the cationic polymerization of p-methoxystyrene in miniemulsion. Basically, the reaction is initiated by the proton (the true initiator), and resultant active centers associate with the pendant aromatic sulfonate moiety (the surfactant and counterion). The bulky ion pairs propagate further until transfer/termination reactions by water occur to generate hydroxyl-terminated oligomers (Scheme 1). All above-mentioned reactions proceed at the particle interface (see Figure 1 for details) providing oligomers with typical Mn e 1000 g/mol. To separate the propagating center from the surfactant counterion, a superacid, trifluoromethanesulfonic acid or triflic acid, was used as an initiator in conjunction with a nonionic surfactant.22 Mn’s of up to 1700 g/mol were produced here, at the expense of the polymerization rate (about ten times slower). Indeed, triflic acid may enter with difficulty the
358
ACCOUNTS OF CHEMICAL RESEARCH
357-367
March 2010
Vol. 43, No. 3
Cationic Polymerization of Vinyl Monomers in Aqueous Media Kostjuk and Ganachaud
SCHEME 1. Proposed Mechanism for p-Methoxystyrene Polymerization in the Presence of DBSA
monomer droplets because of the steric hindrance of the nonionic surfactant molecules, but once done, it apparently goes deeper inside the interface and suffers slower transfer reaction by water than in the INISURF system.
Second Generation: Lanthanide Triflate Cocatalysts Polymerization of p-Methoxystyrene. Sawamoto et al.14 were the first to show that Yb(OTf)3 in conjunction with 1-chloro-1-(p-methoxyphenyl)ethane (pMOS-HCl)14 induces slow cationic polymerization of pMOS at 30 °C in aqueous suspension, while the rate of the reaction considerably increases in the presence of sulfonic acid-based initiators18 (Figure 2a). The use of cationic surfactant (dodecyltrimethylammonium bromide) also allowed the polymerization rate to increase19 (Figure 2a), which is consistent with an increase of the surface area of droplets facilitating the transfer of co-initiator into the organic phase where the polymerization would take place.19 In contrast, anionic surfactant (sodium laurylbenzenesulfonate) or nonionic polymer dispersant (poly(vinyl methyl ether)) retarded the polymerization, as explained by the probable complexation of Yb(OTf)3 with these stabilizers.19 The Mn’s of obtained polymers, regardless of experimental conditions, increased at the beginning of the polymerization (up to ∼30% of conversion) and then leveled off at values between 2500 and 3500 g/mol (Figure 2b) because of chain transfer reactions.14,18,19 Molecular weight distributions (MWD) were typically below 1.5 and did not change with monomer conversion (Figure 2b).14,18,19 Later on, different metal triflates19,20 and tetrafluoroborates20 in conjunction with pMOS-HCl as an initiator were tested by Sawamoto’s team. The reaction rate critically depended on the nature of metal: Yb(OTf)3 > Sn(OTf)2 >
Zn(OTf)2 ≈ Cu(OTf)2.20 Moreover, Sn(OTf)2 was found to be unstable in water and to hydrolyze into inactive hydroxides, while zinc tetrafluoroborate (Zn(BF4)2) induced a very slow polymerization rate.20 Among the various lanthanide triflates tested (Ln(OTf)3; Ln ) Dy, Sm, Gd, Yb, and Nb) together with Sc(OTf)3, the higher activity in suspension polymerization
FIGURE 2. (a) Conversion vs time and (b) Mn and Mw/Mn vs conversion plots for the p-methoxystyrene polymerization with Yb(OTf)3 in aqueous media at 30 °C. Initiator: (1, 3) pMOS-HCl; (2) HOSO2PhNO2. Surfactant: (1, 2) none; (3) dodecyltrimethylammonium bromide.
Vol. 43, No. 3
March 2010
357-367
ACCOUNTS OF CHEMICAL RESEARCH
359
Cationic Polymerization of Vinyl Monomers in Aqueous Media Kostjuk and Ganachaud
SCHEME 2. Mechanism for p-Methoxystyrene Polymerization in the Presence of Yb(OTf)3 Proposed by Sawamoto et al.
(without surfactant) pertained to catalysts with central metal of the smallest ionic radius (Sc(OTf)3, Yb(OTf)3), whereas in emulsion (i.e., using a colloidal stabilizer), Yb(OTf)3 was the most active co-initiator.19 However, the nature of co-initiator did not influence significantly either the Mn or the MWD evolution with conversion.19,20 The analysis of poly(p-methoxystyrene)s chain-ends by 1H NMR spectroscopy was at first revealing the presence of -CH(C6H4OCH3)-Cl ω-terminal group (Fn(Cl) ≈ 0.3-0.5).14,19,20 Later, Sawamoto et al.18 and other research teams16,23-25 showed unambiguously that almost all polymer chains were terminated by hydroxyl groups, originating from chain transfer via water. The low stability of terminal -CH(C6H4OCH3)-Cl groups, even in organic solvents, was also observed by Faust et al.26 Originally, it was proposed14,18-20 that the polymerization proceeds inside monomer/organic solvent droplets via carbocationic species generated from initiator located in organic phase through its reversible activation by Yb(OTf)3. Due to the high hydrophilicity of Yb(OTf)3, only a small amount of co-initiator would be able to penetrate into monomer droplets (Scheme 2). This fact, in the view of the authors,14 explained the necessity to use high amounts of Yb(OTf)3 to reach complete monomer conversion within a reasonable polymerization time. In a recent work, Storey at al.23 reported on the polymerization of pMOS using phosphonic acid initiators [(HO)2P(O)R, R ) Ph or Ph-o,m,p-NO2] and Yb(OTf)3 as co-initiator in conditions similar to those proposed by Sawamoto’s team. It was claimed that polymerization proceeds at the aqueous/organic interface and is not controlled since changing the concentration of initiator did not influence the Mn evolution with conversion.23 The main feature of this initiating system is the possibility to generate poly(p-methoxystyrene)s with high molecular weight (Mn e 20 000 g/mol). An exhaustive mechanistic study of the polymerization of pMOS in the presence of Yb(OTf)3 with special emphasis on the colloidal parameters of the reaction media was performed by Ganachaud et al.17,27 First, conductometric measurements clearly evidenced that ytterbium triflate salt totally dissoci360
ACCOUNTS OF CHEMICAL RESEARCH
357-367
March 2010
Vol. 43, No. 3
ates in water into its ionic counterparts (Yb3+ and CF3SO3-) but does not form nonreactive hydroxides below pH 7, unlike conventional Lewis acids, such as AlCl3 or TiCl4.27 By saturating the water with the catalyst salt, the authors suspected that a few nondissociated Yb(OTf)3 molecules would subsist and reach the interface to co-initiate the polymerization.27 Second, the polymerization was shown to proceed in direct emulsion if an electrosteric-type surfactant such as polyoxyethylene(8) lauryl sulfate sodium was used27 or in inverse (mini)emulsion in the presence of dodecylbenzenesulfonic acid (DBSA).17 In inverse emulsion, polymers with Mn up to 3000 g/mol could be prepared, while in direct emulsion only oligomers with Mn of maximum 1250 g/mol were formed. Moreover, the rate of the polymerization was considerably slower in inverse compared with direct emulsion processes. According to Ganachaud et al.,17,27 the polymerization of p-methoxystyrene in the presence of Yb(OTf)3 proceeds at the interface regardless of the nature of dispersion. Presumably, triflate anion, which arises from the dissociation of Yb(OTf)3, captures a proton and an ytterbium triflate molecule to generate a bulky, strong Brønsted superacid, H+Yb(OTf)4-, which initiates the polymerization. Once a chain is created, propagation is very fast until water transfer/termination reaction occurs to generate hydroxyl-terminated dead oligomeric chains (Scheme 3). In this mechanism, the observed increase of molecular weight at the beginning of the polymerization can be explained by the decrease in the rate of transfer/termination by water due to the increasing hydrophobicity of interface saturated by small oligomers (Figure 1). Otherwise saying, only physicochemical issues govern the molecular weight of the polymers obtained. Polymerization of Vinyl Ethers. The polymerization of isobutyl vinyl ether (IBVE) was initiated by adding a solution of monomer and initiator (adduct of IBVE with HCl, IBVE-HCl) in CCl4 to an aqueous solution of Yb(OTf)3.14 Since Yb(OTf)3 or IBVE-HCl alone did not induce polymerization at all, it was proposed that reaction proceeds via the formation of cationic species through activation of IBVE-HCl by Yb(OTf)3.14 Polymers (61-83% of conversion) with Mn’s scattered around
Cationic Polymerization of Vinyl Monomers in Aqueous Media Kostjuk and Ganachaud
SCHEME 3. Mechanism for p-Methoxystyrene Polymerization in the Presence of Yb(OTf)3 Proposed by Ganachaud et al.
SCHEME 4. Side Reactions of (a) Protonated Monomer and (b) Growing Species during IBVE Polymerization in Aqueous Media
3000-6000 g/mol and broad MWD (Mw/Mn ) 2.0-3.4) were synthesized in, however, a non-reproducible manner. Another study reported the photoinitiated cationic polymerization of vinyl ethers in a mixture of water and acetonitrile, in the presence of Yb(OTf) 3 as co-initiator and diphenyliodonium iodide, which generated monomer-HX adduct during photolysis.28 Oligomers (Mn ) 1800-2600 g/mol) with monomer conversion of 30-50% and narrow MWD (M w/M n ) 1.1-1.3) were produced but again with broad scattering on experimental results.28 A rational explanation for such random polymerization results lies in the irreversible consumption of protonated monomer29 or degradation of carbocationic growing species,14,30 because of their high reactivity toward water and the low stability of hemiacetals formed in such aqueous environment (Scheme 4).
Third Generation: Lewis Acid-Surfactant Combined Catalysts (LASC) Since all polymerizations of pMOS discussed above proceed at the interface, the “critical DP” effect limits the molecular weight at values around 1000 g/mol (direct emulsion) or 3000 g/mol (suspension, inverse emulsion), respectively. To produce longer polymer chains, it is necessary to find conditions where the polymerization locus could be displaced from the interface to the core of the monomer droplets.16 For this purpose, a complex of Lewis acid with surfactant, that is, socalled Lewis acid-surfactant combined catalyst (LASC) was investigated in miniemulsion polymerization of pMOS.24 Typically, the tris(dodecyl sulfate)ytterbium LASC was recovered after precipitation in an aqueous mixture of Yb(OTf)3 and dodecyl sulfate sodium salt (1:3).24 Unfortunately, the authors undoubtedly showed that this particular LASC stayed at the
interface and did not affect the polymerization process, that is, it acts only as a surfactant together with dodecyl sulfate sodium salt.24 Later on, another LASC was (unwittingly) generated under aqueous miniemulsion conditions, by in situ coupling between an electrosteric surfactant and ytterbium but (purposely) associated with a weak organic acid as an initiator to generate large molar mass polymer chains (Scheme 5a).25 According to the authors,25 only small oligomers were formed at the beginning of the reaction by interfacial polymerization. In strong contrast to previous studies however, high molecular weight polymers (Mn up to 40000 g/mol) started to appear after ca. 100 h. At this stage, the reaction media also changed from a two-phase mixture to a fine emulsion (Figure 3). It was proposed that, in this second stage, the polymerization proceeds inside the monomer droplets through the following steps: (i) in situ LASC formation; (ii) association with the weak acid to generate a bulky superacid and initiate the polymerization; (iii) fast propagation; (iv) (moderate) transfer due to the low content of water inside the droplets, thus allowing chains to grow significantly.25 Since no variation of molecular weight with conversion was observed during the reaction, it was concluded that the polymerization was not controlled under these conditions.
Fourth Generation: BF3 · OEt2 Cocatalyst Along with Yb(OTf)3, another strong Lewis acid, BF3 · OEt2, was tested as a possible water-tolerant co-initiator for the cationic polymerization of styrene and its derivatives.31-36 We describe here mainly the results obtained in “wet” solution, that is, where (an excess of) water (against Lewis acid) was introduced as a polymerization controlling agent. Vol. 43, No. 3
March 2010
357-367
ACCOUNTS OF CHEMICAL RESEARCH
361
Cationic Polymerization of Vinyl Monomers in Aqueous Media Kostjuk and Ganachaud
SCHEME 5. (a) LASC Formation in Water and Superacid Generation in the Organic Phase and (b) Proposed Cationic Polymerization Mechanism inside the Monomer Droplets
p-Hydroxystyrene and p-Alkoxystyrenes Polymerization. The cationic polymerization of such active monomers as p-hydroxystyrene (pHS)31 and p-alkoxystyrenes32 was first studied by Sawamoto et al. using different adducts of p-methoxystyrene as initiators (Scheme 6) and BF3 · OEt2 as a co-initiator in acetonitrile or its mixture with CH2Cl2. It was found that alcohol (4) or ether pMOS-derived initiators (3) induced faster polymerization in contrast to chloride (1) and acetate (2) functionalized initiators, typically used in common metal chloride-mediated controlled cationic polymerizations. Importantly, the addition of excess water toward Lewis acid is necessary to
FIGURE 3. (a) Scanning and (b) transmission electron microscopy photographs of one example of final polymer dispersion.
362
ACCOUNTS OF CHEMICAL RESEARCH
357-367
March 2010
Vol. 43, No. 3
SCHEME 6. p-Hydroxystyrene Polymerization Co-initiated by BF3OEt2 in Acetonitrile and in the Presence of Excess Water
control pHS polymerization with the 4/BF3OEt2 initiating system, advantageously without the constraint of protecting the hydroxy groups.31 Increasing the water concentration in the system led to more controlled polymerization in terms of Mn and MWD evolution with conversion, however at the expense of the reaction rate (Figure 4): the optimal [H2O]/[BF3OEt2] ratio was found to be 100 (CH3CN/H2O 250:1 v/v) (Figure 4). The living nature of pHS polymerization was confirmed by a monomer postaddition experiment: the Mn increased from 3000 to 6000 g/mol and MWD stayed relatively narrow (Mw/Mn < 1.4). Basically, poly(p-hydroxystyrene)s with molec-
Cationic Polymerization of Vinyl Monomers in Aqueous Media Kostjuk and Ganachaud
FIGURE 4. Effects of Water on the pHS Polymerization with 4/BF3OEt2 in Acetonitrile at -15 °C: [pHS] ) 0.20 M; [4] ) 4.0 mM; [BF3OEt2] ) 2.0 mM.
ular weight up to 15000 g/mol (Mw/Mn∼2.0) and 6000 g/mol (Mw/Mn < 1.4) could be synthesized by pHS polymerization with the 4/BF3OEt2 initiating system at [H2O]/[BF3OEt2] ratio of 6 and 100, respectively.31 The same initiating system was also successfully applied to the polymerization of p-alkoxystyrenes (p-methoxystyrene, p-tert-butoxystyrene (tBOS)).32 The polymerization rate strongly depended on the nature of the para substituent: tBOS